What an 89 mg/dL Blood Sugar Reading Tells You

89 mg/dL (4.9 mmol/L) is a normal glucose level. Fasting, it sits comfortably below the 100 mg/dL threshold that defines impaired fasting glucose, and well above the 70 mg/dL level-1 hypoglycemia alert threshold. Post-meal or random, 89 is unremarkable. If you are not on insulin or a sulfonylurea, an 89 is a non-event. The interesting question is what a single 89 lets you conclude, and the answer is: less than you think. The number has an error bar, a context, and a distribution behind it, and all three matter more than the point estimate.

What the number means depends on where it came from

Three devices can all report 89 and mean different things.

A venous plasma glucose from a lab is the reference method, and it has a preanalytical failure mode people forget: glycolysis in the tube. Red and white cells keep consuming glucose after the draw. In an uncentrifuged sample at room temperature, glucose falls on the order of 5 to 7 percent per hour, so a true 95 can be reported as 89 if the tube sat for an hour before spinning. Sodium fluoride slows this but does not stop it quickly enough to help in the first hour. Citrate-buffered tubes do better, and laboratory guidelines specifically address this handling problem because it shifts population-level diagnostic classification ¹. If your fasting glucose is bouncing between 89 and 99 across draws at the same lab, ask about tube type and time-to-centrifugation before you interpret the trend.

A fingerstick meter is held to ISO 15197:2013, which requires 95 percent of results within ±15 mg/dL of the reference below 100 mg/dL. At 89, that permits a true value anywhere from 74 to 104 and still counts as a passing meter. Hematocrit, cold hands, and residual sugar on the fingertip all push further. A second strip from the same drop routinely differs by 5 to 8 mg/dL.

A CGM reports interstitial fluid glucose, not blood glucose, and converts it through a calibration model. Interstitial glucose tracks plasma with a physiological lag of roughly 5 to 15 minutes, longer when glucose is moving fast ². Correlation between interstitial and venous plasma is good in steady state and degrades during rapid excursions ². Current sensors run a mean absolute relative difference around 8 to 10 percent, so an 89 on a Dexcom or Libre corresponds to a plasma value with a standard deviation of roughly 7 to 9 mg/dL around it. Overnight, a CGM 89 can also be a compression low from sleeping on the sensor, which shows up as a smooth dip with a fast recovery on rolling off.

None of this makes 89 a bad reading. It means you should treat any single glucose value as a draw from a distribution. A lone reading is not a fixed fact.

A single 89 does not give you an A1c

The common follow-up question is what A1c corresponds to 89. The standard regression from continuous monitoring data relates mean glucose to A1c as mean glucose (mg/dL) ≈ 28.7 × A1c − 46.7. Inverting, a 24-hour mean glucose of 89 maps to an A1c near 4.7 percent.

Two caveats make that arithmetic misleading if you apply it to one fingerstick. First, the input is a multi-week mean. The formula does not take a fasting point. Fasting glucose is usually the lowest part of your daily curve, so plugging a fasting 89 into the formula understates your mean and therefore your A1c. Someone with a fasting 89 and large post-meal excursions can have a 24-hour mean near 110. Second, the mapping has real between-person scatter: the same mean glucose produces A1c values spanning roughly half a percentage point across individuals, driven by red cell lifespan and glycation rate. Hemoglobin variants, recent blood loss, iron deficiency, and hemolysis all shift A1c independent of glucose, which is why laboratory guidance treats A1c and glucose as complementary rather than interchangeable measurements ¹.

If you want your mean glucose, measure it. Wear a CGM.

What to compute instead of staring at one number

Fourteen days of CGM at 5-minute resolution gives you about 4,000 readings. That is enough to estimate a mean, a variability coefficient, and an overnight nadir with real precision. Wear-time completeness matters: in pediatric type 1 cohorts, wearing the sensor at least 90 percent of the time produced measurably tighter glycemic outcomes than 70 to 89 percent wear, so partial datasets both bias the statistics and degrade the behavior they reflect ³.

Dexcom exports a CSV from Clarity with Timestamp (YYYY-MM-DDThh:mm:ss) and Glucose Value (mg/dL) columns, with Low/High strings for out-of-range clamps. Libre exports a CSV with a Record Type column where 0 is historic and 1 is scan. Load and clean before you compute anything:

import pandas as pd, numpy as np

df = pd.read_csv("clarity_export.csv")
df = df[df["Event Type"] == "EGV"].copy()
df["t"] = pd.to_datetime(df["Timestamp (YYYY-MM-DDThh:mm:ss)"])
g = (df["Glucose Value (mg/dL)"]
       .replace({"Low": 39, "High": 401})
       .astype(float))
s = pd.Series(g.values, index=df["t"]).sort_index()

# wear-time completeness at 5-min cadence
expected = (s.index[-1] - s.index[0]) / pd.Timedelta("5min")
print(f"completeness: {len(s)/expected:.1%}")

mean = s.mean()
cv   = s.std(ddof=1) / mean
tir  = ((s >= 70) & (s <= 140)).mean()
night = s.between_time("00:00", "06:00")
print(f"mean {mean:.1f}  CV {cv:.1%}  TIR70-140 {tir:.1%}  "
      f"night p5 {np.percentile(night, 5):.0f}")

The four numbers worth tracking, for a person without diabetes:

• Mean glucose. A 14-day mean in the high 80s to mid 90s is typical for a metabolically healthy adult.
• Coefficient of variation. Below 36 percent is the standard stability threshold. Healthy non-diabetic CV usually lands near 15 to 20 percent.
• Percent of readings in 70 to 140. A tighter band than the clinical 70 to 180 target, and more informative if you are not diabetic.
• Overnight 5th percentile. Nocturnal values in the 70s and low 80s are common and not concerning by themselves. Repeated sub-70 stretches are worth showing to a clinician.

Then add per-meal analysis: peak, time-to-peak, and 2-hour delta after the same meal eaten on different days. The within-person day-to-day variance for an identical meal is large, so a single test of a food tells you almost nothing. Repeat it three times before you believe the result. CGM was built and validated for exactly this kind of pattern detection rather than for single-point accuracy ⁴.

When 89 does matter

If you take insulin or an insulin secretagogue, the context changes. The ADA and Endocrine Society workgroup set the glucose alert value at 70 mg/dL, with clinically significant hypoglycemia below 54 and severe hypoglycemia defined by needing another person’s help rather than by a number ⁵. An 89 in that setting is not low. A downward trend arrow on a CGM at 89 with insulin on board is still a different situation from a flat 89. Exercise shifts this as well: the type 1 exercise consensus describes glucose falling during and for hours after aerobic activity, which is why pre-exercise numbers get read together with trend and insulin timing ⁶. How you manage any of that is a conversation with your diabetes clinician. A blog post is not the place to work it out.

The other case where a fasting number in the 80s to low 100s carries genetic information is heterozygous glucokinase variants. These raise the beta cell’s glucose set point and produce a lifelong, mildly elevated and stable fasting glucose from a primary secretory defect ⁷. That presents as a persistent fasting value in the 100 to 145 range rather than 89, but it is a good illustration of why the fasting set point is partly a coded parameter and why a whole-genome sequence is a useful companion to a glucose trace.

Questions people also ask

Is 89 okay for blood sugar? Yes. It falls inside the normal fasting range (below 100 mg/dL) and above the 70 mg/dL low threshold used in diabetes care ⁵.

Is 89 hypoglycemia? No. The standard alert value for hypoglycemia is 70 mg/dL, and clinically significant hypoglycemia is defined at below 54 mg/dL ⁵. Some people feel adrenergic symptoms at higher glucose values, especially after a period of running high, but symptoms at 89 are not hypoglycemia by the measurement.

Is 89 too low for a diabetic? It is within the usual pre-meal target band used in diabetes management. Whether it is comfortable for a given person depends on their medications and their hypoglycemia awareness, and that assessment belongs with their clinician.

What is your A1c if your blood sugar is 89? A 24-hour mean glucose of 89 mg/dL corresponds to an A1c near 4.7 percent using the standard regression, but a single fasting 89 does not establish your mean, and A1c varies across people at the same mean glucose ¹.

What should I do if my blood sugar is 89? Nothing, if you are not on glucose-lowering medication. If you want more information than a single reading offers, wear a CGM for 14 days with high completeness and compute the mean, CV, and overnight nadir ³.

Footnotes

1. David B. Sacks, Mark A. Arnold, George L. Bakris, et al. Guidelines and Recommendations for Laboratory Analysis in the Diagnosis and Management of Diabetes Mellitus. Diabetes Care, 2023. https://doi.org/10.2337/dci23-0036 ↩ ↩² ↩³

2. Young Ha Baek, Heung Yong Jin, Kyung Ae Lee, et al. The Correlation and Accuracy of Glucose Levels between Interstitial Fluid and Venous Plasma by Continuous Glucose Monitoring System. Korean Diabetes Journal, 2010. https://doi.org/10.4093/kdj.2010.34.6.350 ↩ ↩²

3. Alzbeta Šantová, Vít Neuman, Lukas Plachy, et al. The Longer, the Better: Continuous Glucose Monitoring Use for ≥90% Is Superior to 70%–89% in Achieving Tighter Glycemic Outcomes in Children with Type 1 Diabetes. Diabetes Technology & Therapeutics, 2025. https://doi.org/10.1089/dia.2024.0472 ↩ ↩²

4. Hun‐Sung Kim, Jeong‐Ah Shin, Jin‐Sun Chang, et al. Continuous glucose monitoring: current clinical use. Diabetes/Metabolism Research and Reviews, 2012. https://doi.org/10.1002/dmrr.2346 ↩

5. Elizabeth R. Seaquist, John E. Anderson, Belinda P. Childs, et al. Hypoglycemia and Diabetes: A Report of a Workgroup of the American Diabetes Association and The Endocrine Society. Diabetes Care, 2013. https://doi.org/10.2337/dc12-2480 ↩ ↩² ↩³

6. Michael C. Riddell, Ian Gallen, Carmel E. Smart, et al. Exercise management in type 1 diabetes: a consensus statement. The Lancet Diabetes & Endocrinology, 2017. https://doi.org/10.1016/s2213-8587(17)30014-1 ↩

7. Gilberto Velho. Primary pancreatic beta-cell secretory defect caused by mutations in glucokinase gene in kindreds of maturity onset diabetes of the young. The Lancet, 1992. https://doi.org/10.1016/0140-6736(92)91768-4 ↩